Electromagnetic scattering characteristic analysis method for parallel acceleration of characteristic basis function algorithm based on GPU

Abstract

The invention discloses an electromagnetic scattering characteristic analysis method for parallel acceleration of a characteristic basis function algorithm based on a GPU. The method comprises the following steps: carrying out three-dimensional modeling on a to-be-detected target; carrying out mesh generation on the model of the to-be-detected target based on an RWG basis function; optimizing thecharacteristic basis function algorithm by using a discrete Galerkin method based on the mesh generation information; carrying out parallel acceleration on the characteristic basis function algorithmby utilizing the GPU; and taking the mesh generation information as the input of the characteristic basis function algorithm after parallel acceleration, and outputting electromagnetic scattering characteristics of the to-be-detected target. According to the method, the discrete Galerkin method is used for optimizing the characteristic basis function algorithm, and the GPU is used for carrying outparallel acceleration on the characteristic basis function algorithm, so that the electromagnetic scattering problem in the separation process of the electrically large-size target and the complex target can be solved, and calculation has high accuracy and high efficiency.

Description

technical field [0001] The invention belongs to the field of computational electromagnetics, in particular to an electromagnetic scattering characteristic analysis method based on a GPU parallel accelerated characteristic basis function algorithm. Background technique [0002] With the continuous development of society, the progress of science and technology is changing with each passing day. In the field of engineering applications, people urgently need to study the electromagnetic characteristics of complex or electrically large targets. Typical applications include: target stealth and anti-stealth design, radar system design and target recognition, etc. The research methods mainly include measurement method and calculation method. The measurement method plays an important and irreplaceable role in testing the electromagnetic characteristics of the target, but it has the disadvantages of high test cost and long test cycle. With the rapid improvement of computer performan...

AI Technical Summary

Problems solved by technology

However, such methods need to artificially introduce boundaries outside the solution domain and have a large amount of unknowns when the solution domain is divided into volumes. In addition, they have the inherent disadvantages of differential methods: the grid dispersion error is large, and as the solution domain increases, its The greater the accumulation of dispersion errors, so this type of method is not suitable for solving the problem of electromagnetic characteristics of electrically large targets

However, the impedance matrix obtained by the MoM method is full, and its storage complexity is O(N 2 ), the computational complexity is O(N 2 ) or O(N 3 ), so the traditional MoM method can only analyze the electromagnetic characteristics of electrically small targets

[0005] Fast algorithms based on the method of moments have developed rapidly, such as the fast multipole method (FMM) and the multilayer fast multipole method (MLFMA). These methods do not explicitly fill in the impedance matrix of the far-field mutual coupling when analyzing electrically large targets. The iterative method must be used to solve the linear equation system. When the target is more complex, the iterative convergence is usually very slow. When solving the electromagnetic response problem of the target under multiple incidents, the method of iterative solver further shows disadvantages

Images